Biggest problem of file sharing is the little upload bandwith.

Is it possible to address the same upload to different people?

I have a file "xyz" that 5 guys want download from me, i have a max upload bandwith of 15 Kb/s: can the file be addressed to all at 15Kb/s?

Effective upload is 15 Kb/s but each one of them will receive 15 Kb/s so virtually a total of 75 Kb/s....

Something like streaming, i sent a person some bytes and other people interested in can have them at the same time.

Something like tell the program/protocol "sent filexyz to 210.11.23.32 and to 213.34.54.66 and to...."


Sorry for poor english, i hope someone had understand the post; and sorry also because i'm not a coder so i can't help in programming

I think you got something wrong...
If you send the file to 5 people you divide 15 kbps with 5.
You cant send a file to someone with no bandwidth.

You could use multicasting, but there are two problems with that.

First is that most ISPs block multicast traffic (you send a byte, multiple recipients receive that same byte). Because the entire path between you (the uploader) and them (the downloaders) needs to be multicast enabled, you undoubtly run into a few points (ISPs) along that path that don't have multicasting enabled.

The second problem is that multicasting is UDP based. It stands for User Datagram Protocol, but you could freely exchange that with "Unreliable Datagram Protocol" because UDP packets are not guaranteed to arrive at recipient. So a few recipient could end up swiss cheese - a file with lots of missing bytes. You could correct that by re-requesting a certain part of the file, but the cycle can continue over and over - thus eliminating what you were supposed to gain (speed).

With the advent of HUGE, some of these obstacles could be eliminated. If User 1 is downloading file XYZ and User 2 is requesting the same file, through HUGE you could tell User 2 to ask User 1 (who should already have portions of that file at that point). However, this would mean that User 1 must be able to upload a file it is currently downloading AND is sharing files...

Sorry for my poor english

It's true: actually i divide the total ammount of bandwith for the number of upload people so it's like u tell.

But my idea is to "change gnutella protocol": i'd like to have the possibility of a streaming from one peer to multiple peers receive it.

I see in a live streaming event that the bandwith in upload from a server is for example of 30KB/s and 100 men can receive it but the server doesn't use a bandwith of 3000 KB/s, use only 30KB/s!

Data trasmitted is the same but i address it to different people at the same time: I tell send the file XXX to different Ip addresses but using the same ammount of bandwith of only one upload, so bandwith will not be divide for number of people who download the same file as is actually!

I don't know if it's possible, i don't know if i explain clear my idea and i don't know if this is the right place to explain my idea: please tell me about

Thanks for your attention

Maybe should be possible implement this feature for people who had a ISP that permit this type of multicast traffic; that should be a little improvement of network.
I think that also if there's 1% of users who can use multicasting the upload should be best.

If the problem of not guaranteed arrive of packet is due to receiver bandwith, should be possible splitting the uploader velocity in little piece that could be surely received.....

For example i have 15 KB/s of bandwith but i divide it little pieces of 2KB/s so also people with slow or difficulties connections can download it surely. If downloader can't guarantee such bandwith he'll be excluded from the streaming and will download the file from other source......

So HUGE can't resolve the problem of uploading bandwith....


Thanks for your attention and can u tell me about where i can find documentation about HUGE, UDP and so on?

I have another "idea": instead of multicasting isn't possible to "sniffing" low bandwith comunication between two peers?

Peer A transmits only to Peer B and a peer C, interested at the same file "intercept" the data transfert from the peers A and B.

The transfert could be at low speed to avoid too much packet are loss, and eventually correction of the errors can be made at the end of transmission of file....

So Peer A don't need to use multicasting/UDB to address the same file to different peers.....

Sorry if this idea is crazy or like in a science-fiction but i don't know about protocols and programming